RPA and iowa

[Lloyd Zusman]

This question has to do both with RPA and with iowa.

I have recently installed iowa via RPA as suggested on the iowa web
site (on this page: http://enigo.com/projects/iowa/index.html).

Now, I'm going through the iowa tutorial, but I can't find the WEBrick
example. The tutorial says to do the following:

cd examples/webrick/iowa
ruby webrick_iowa.rb

However, I don't have a clue as to where RPA put the
"examples/webrick/iowa" directory, or for that matter, if it even
installed that directory at all.

Can anyone point me to where that directory lives when iowa is installed
via RPA?

If that directory doesn't exist, I know that I can get it by downloading
and unpacking the iowa tarball from SourceForge. However, that obviates
the need to use RPA in the first place.

And if indeed SourceForge is a better choice for iowa downloads, what
will happen from now on if I give up on RPA? In using it for this iowa
download, it created a package tree that contains the installations of
iowa and all its dependencies, some of which are upgrades of packages
that I had previously installed outside of RPA. If I stop using RPA,
will I have to remove the RPA package tree in order to keep my installs
in sync? If so, will I have to re-install everything that RPA has
recently put under its tree?

Thanks in advance for any insights you might be able to offer about all
of this.

 

[Lloyd Zusman]

CORRECTION: in the message below, please change all occurrences of
the string "SourceForge" to "RubyForge".

 

[Mauricio Fernández]

However, I don't have a clue as to where RPA put the
"examples/webrick/iowa" directory, or for that matter, if it even
installed that directory at all.

Can anyone point me to where that directory lives when iowa is installed
via RPA?

The documentation is installed under $prefix/share/doc/rpa0.0; this is
followed consistently for all ports.

You can get the list of files associated to the iowa package with
rpa info -Dfiles -x iowa

And if indeed SourceForge is a better choice for iowa downloads, what
will happen from now on if I give up on RPA? In using it for this iowa
download, it created a package tree that contains the installations of
iowa and all its dependencies, some of which are upgrades of packages
that I had previously installed outside of RPA. If I stop using RPA,
will I have to remove the RPA package tree in order to keep my installs
in sync?

There is no need to remove the packages: note that rpa-base installs
the .rb files under $prefix/lib/ruby/site_ruby/1.8 (i.e. the standard
library path), and they are used with the normal Kernel.require.
The files under $prefix/lib/ruby/rpa0.0 would only be used if you
require 'rpa'

If you don't want to use it anymore, you can remove rpa-base itself with
rpa remove rpa-base
This is not atomic for obvious reasons
The scripts that use "require 'rpa'", if you have written any, would
stop working. If this is a problem and you still want to remove rpa-base,
you can create a dummy rpa.rb like

RPA = Struct.newversion).new

If so, will I have to re-install everything that RPA has
recently put under its tree?

No. You can install over files "owned" by rpa-base. The files in the
standard site_ruby dir will be used preferentially, unless you
explicitly require 'rpa' .

 

[Lloyd Zusman]

The documentation is installed under $prefix/share/doc/rpa0.0; this is
followed consistently for all ports.

Aha! Yes, what I'm looking for is there.

You can get the list of files associated to the iowa package with
rpa info -Dfiles -x iowa

Thanks, this is very useful.

There is no need to remove the packages: note that rpa-base installs
the .rb files under $prefix/lib/ruby/site_ruby/1.8 (i.e. the standard
library path), and they are used with the normal Kernel.require.
The files under $prefix/lib/ruby/rpa0.0 would only be used if you
require 'rpa'

If you don't want to use it anymore, you can remove rpa-base itself with
rpa remove rpa-base
This is not atomic for obvious reasons
The scripts that use "require 'rpa'", if you have written any, would
stop working. If this is a problem and you still want to remove rpa-base,
you can create a dummy rpa.rb like

RPA = Struct.newversion).new

OK. This clears things up for me.

No. You can install over files "owned" by rpa-base. The files in the
standard site_ruby dir will be used preferentially, unless you
explicitly require 'rpa' .

Thanks for all your help!

 

[Mauricio Fernández]

Aha! Yes, what I'm looking for is there.

All the useful information from the upstream tarball should be preserved
in the RPA port. If not, you can consider it a bug (see below).

Forgot to say that those paths are relative to the $prefix you specified
when installing rpa-base.

If you have any problems with the iowa (or another) RPA port or some
idea to make it better, feel free to issue a bug report at

http://rubyforge.org/tracker/?atid=1087&group_id=265&func=browse

under the "Ports" category.

 

[leon breedt]

http://rubyforge.org/tracker/?atid=1087&group_id=265&func=browse

Erm, who wants to write some RewriteRules using a RewriteMap to lookup
project IDs for RubyForge so we can make URLs like the above one
pretty?

Possible replacement candidates:

http://rubyforge.org/tracker/rpa-base
http://rubyforge.org/projects/rpa-base/tracker

Or even just a redirect from a friendly URL to the full one would
serve the purpose of making it easy to give out the URL to people, as
well as future proof it.

I know I won't remember that URL should I need to report bugs in the future

Leon

 

[Tom Copeland]

Erm, who wants to write some RewriteRules using a RewriteMap to lookup
project IDs for RubyForge so we can make URLs like the above one
pretty?

Possible replacement candidates:

http://rubyforge.org/tracker/rpa-base
http://rubyforge.org/projects/rpa-base/tracker

.Or even just a redirect from a friendly URL to the full one would
serve the purpose of making it easy to give out the URL to people, as
well as future proof it.

I know I won't remember that URL should I need to report bugs in the future

Could be done... it's a bit tricky, though, because some projects have
disabled the trackers and thus wouldn't have a destination for a URL
like that. I kind of feel like most folks just go to the project page
and start clicking from there, although I could be wrong...

Yours,

Tom